What are Continuous Delivery and Continuous Deployment?

Traditionally, the development of software is divided into three different phases - programming individual components (Development), integrating all necessary components (Integration), and finally delivering or deploying a finished product to end customers (Delivery). In the past, these three stages were completed chronologically and linearly, leading to significant problems in transitioning from one stage to the next. For this reason, modern software development concepts advocate for a close integration known as Continuous Integration, Continuous Delivery, and Continuous Deployment. These are known as the CI/CD model, where CD can stand for Continuous Deployment, Continuous Delivery, or both simultaneously.

Contents:
What are the basic principles and definition of Continuous Delivery?
What tasks does Continuous Delivery encompass?
What is behind Continuous Deployment?
What are the advantages of Continuous Delivery?
Which tools assist in implementing Continuous Delivery?
 

What are the basic principles and definition of Continuous Delivery?

Unlike other products, software development in most cases is not a static process that leads to a permanently market-ready result once completed. Instead, it usually requires regular updates that introduce new features or address security vulnerabilities that have been identified. Even if this is not the case, the development process requires the recurring creation of alpha and beta versions, which are at least delivered to a fixed group of testers, other developers, and potentially interested members of the public.

Instead of viewing the individual phases as encapsulated development steps, complex projects adopt an alternative approach that intertwines the different stages through Continuous Integration and continuous delivery. Continuous Delivery represents a collection of tools, processes, and methods that complement the delivery of software in addition to traditional methods such as selling physical media or providing it as a free or paid download. Certain specific requirements must be met for this:

• Comprehensive version control for the source code
• Continuous integration as the basis for rapid software development
• Automation of recurring processes such as compiling, testing, and package creation through suitable tools
• Continuous delivery through central platforms like GitLab
• Regular feedback between different teams and (test) users
• Short development cycles with regular builds for practical use

Therefore, the definition of continuous delivery encompasses the entire development process, which is not linear but cyclical in a tight and continuous loop of progress, automated builds and tests, quality control, and feedback to the preceding stages. It typically includes the insertion of developers' work carried out within the scope of Continuous Integration, flowing into the finished product in short intervals in the form of Nightly Builds, allowing for almost immediate error checking.

What tasks does continuous delivery involve?

Continuous delivery is the second step after Continuous Integration and aims to rigorously and continuously test and optimise the quality of software. It does not deal with the actual development such as writing code, but rather represents a method for efficient and comprehensive quality control during the creation and further development. To achieve its goal, projects are in an interim phase, the Continuous Delivery Pipeline, where they are checked manually and automated using suitable tools. The tasks involve several steps that are carried out in chronological order or as parallel as possible:

• Changes or extensions in the code as part of Continuous Integration
• Checking the version, compiling, and creating executable bundles such as Nightly Builds or Beta versions (Commit Stage)
• Automated tests for performance, stability, integration, functionality, and system compatibility (Acceptance Test Stage)
• Transmission of logs from the last stage, error correction, and bug fixing (Feedback Stage)
• Manual tests in one or more environments and checking individual relevant components (Manual Test Stage)
• Continuous delivery via own or external infrastructure (Delivery Stage)

Delivery takes place through different channels and to varying extents, for example in the form of physical media or as a public or protected download for groups such as customers and developers. An alternative and more advanced method is Continuous Deployment, which is preferred for large or security-critical applications. The tasks of continuous delivery encompass certain areas of Continuous Integration but expand it with an additional aspect that includes the release of products.

What is behind Continuous Deployment?

Although in publications outside the specialist press and in general usage, a specific distinction between Continuous Delivery and Continuous Deployment is sometimes not made, there is a significant technical difference. With Continuous Deployment, development takes a further step and automates not only code integration and testing but also the continuous delivery - thus forming a optional extension of the process within Continuous Delivery, which enhances security by promptly installing updates for improving performance or functionality, as well as addressing security vulnerabilities on productive systems. Distribution is carried out through various means such as automatic updates, which are delivered either via their own servers or in the form of specialised Infrastructure as a Service (IaaS). Additionally, there are increasingly platforms like GitLab that explicitly integrate Continuous Deployment as part of their concept alongside Continuous Integration and Continuous Delivery. Furthermore, there are providers offering CI/CD together with complete development environments as Software as a Service (SaaS).

What are the benefits of continuous delivery?

By combining Continuous Integration and continuous delivery, almost all important steps in software development are linked together and the entire process is made cyclic instead of linear. This results in numerous benefits for both developers and customers:

• Rapid identification of code errors facilitates debugging and prevents misdevelopments
• Continuous and automated testing speeds up development and reduces costs
• Increased efficiency by eliminating manual builds and releases
• Direct and prompt feedback enhances developers' work and highlights individual weaknesses
• Faster, more frequent, and higher-quality releases for customers
• Fully automated installation on productive end systems through Continuous Deployment
• Concentration of all resources on essential processes avoids downtime and extra work

Due to the numerous benefits, more and more companies are transitioning to consistently rely on CI/CD for their projects and implementing the corresponding procedures. In practice, this often means a higher investment in, for example, integration servers and test environments, which, however, usually pay off relatively quickly due to the improved and faster workflow. Some conditions for successful implementation include appropriate team structuring and optimisation of automated tests - as errors in these areas can have a significant impact on the stability, security, performance, and reliability of the software.

Which tools help with the implementation of Continuous Delivery?

Establishing a complete and optimised environment for Continuous Integration and continuous deployment can be particularly challenging and time-consuming for small and medium-sized enterprises (SMEs). Essential in any case are professional tools such as the web application Jenkins, known for its clear and user-friendly interface and functionality. An alternative to building your own infrastructure is to opt for pre-built solutions from professional providers such as GitLab or Codeship, which offer CI/CD as a service in the form of SaaS or IaaS. Customers can choose from several packages with varying levels of functionality or a customised environment, which may also include specialised development environments as a central hub for the involved programmers, depending on the service provider's offerings. Alternatively, with the appropriate expertise and resources, it is also possible to build your own customised platform on a powerful Virtual Private Server, Cloud Server, or Dedicated Server, suitable for a single project or as a permanent environment. For example, GitLab, as open-source software, can be installed freely and without restrictions on user numbers and functionality on a self-managed infrastructure or on Managed Servers.

Photo: Gerd Altmann on Pixabay